In This Week’s SuperDataScience Newsletter: Microsoft to Eliminate Face Analysis in Move Towards ‘Responsible AI.’ Instagram is Testing an AI Face-Scanning Tool for Age Verification. China Achieves ‘Brain-Scale’ AI with Latest Supercomputer. Big Tech Models are Challenged by Open-Source Language AI. Data Science Facebook Groups You Should Join Today. Cheers,
- The SuperDataScience Team P.S. Have friends and colleagues who could benefit from these weekly updates? Send them to [this link]( to subscribe to the Data Science Insider. --------------------------------------------------------------- [Microsoft to Eliminate Face Analysis for ‘Responsible AI’]( brief: Microsoft has announced that in a move towards a more “responsible AI standard” they will no longer allow its technology to be used in certain ways, such as to infer emotion, gender, or age using facial recognition technology. A product manager at Microsoft, Sarah Bird, commented: “We collaborated with internal and external researchers to understand the limitations and potential benefits of this technology and navigate the tradeoffs. In the case of emotion classification specifically, these efforts raised important questions about privacy, the lack of consensus on a definition of ‘emotions’, and the inability to generalise the linkage between facial expression and emotional state across use cases.” Azure Face service is one of the features which will change with companies who use the tool (such as Uber) now having to apply to Microsoft and prove they match their AI standards in order to be able to use it. Why this is important: Ethical AI is a hot-topic conversation and Microsoft seems to be taking the issue seriously. The latest announcements don’t mean that they are scrapping emotion recognition technology entirely, but they are shifting towards using it in a more considered and responsible manner. [Click here to learn more!]( [Instagram Tests AI Face-Scanning Tool for Age Verification]( brief: Instagram has announced that it is trialing several new means of verifying users' ages including a face-scanning AI tool, having mutual friends verify their age, or uploading an ID. The move is intended to crack down on users who fall below Instagram’s thirteen-year-old age limit but will initially only look to verify users who are eighteen or older. If users choose to use AI to verify their age, then a selfie will be sent to a third-party company, Yoti, which uses ML to make an estimation. Yoti uses various facial signals to estimate a target’s age but claims to not know exactly which criteria pictures are assessed against. Yoti’s technology has previously been approved for use by the UK government and German digital regulators and they are well-known in the online age and ID verification sphere. They claim to not receive any data apart from the photo which it doesn’t retain. Why this is important: It is interesting that this announcement comes in the same week as Microsoft reveals that it is moving away from facial scanning technology. Some of the criticisms about the ethics of facial scanning technology are upheld in the data about Yoti’s accuracy, which is less precise when it comes to analysing images of women and those with darker skin tones. [Click here to read on!]( [China Achieves ‘Brain-Scale’ AI with Supercomputer]( In brief: Chinese scientists revealed a new supercomputer, the Newest Generation Sunway supercomputer, which they claim is so fast that it has successfully run an AI model as sophisticated as a human brain. The Chinese team trained an AI model called bagualu (which means alchemist’s pot), using Sunway, with 174 trillion parameters. For the first time ever, this was said to rival the number of synapses in the brain; however, as this Interesting Engineering article points out, it is incredibly difficult to map the exact number of synapses in a brain, with some estimates suggest the human brain contains up to 1,000 trillion synapses. According to researchers, the Sunway supercomputer has a speed of a billion operations per second or 5.3 floating-point operations per second. It also has 37 million CPU cores and nine petabytes of memory. Researchers also claim the 96,000 semi-independent computer systems resemble the power of a human brain. Why this is important: The achievement puts the Newest Generation Sunway supercomputer on par with Frontier, the latest machine built by the US Department of Energy, which was named the world’s most powerful earlier this month. The supercomputer race is back on! [Click here to discover more!]( [Big Tech Models Challenged by Open-Source Language AI]( In brief: Academic researchers have attempted to break what they consider to be “big tech’s stranglehold on natural-language processing and reduce its harms.” The language model, called BLOOM, has been trained with US$7-million-worth of publicly funded computing time and will rival in scale those made by Google and OpenAI, despite being open-source and the first model of its scale to be multilingual. Researchers, both those involved in the development and those who are independent, have praised the AI’s development and potential. Thomas Wolf, lead on the initiative and co-founder of Hugging Face, a company that hosts an open-source platform for AI models and data sets, stated: “We think that access to the model is an essential step to do responsible machine learning,” and independent Connor Leahy said: “It was long overdue that this technology diffused into the open-source world, and this is quite an interesting way for it to have happened.” Why this is important: As regular readers of the SuperDataScience newsletter will know, models that recognise and generate language are becoming increasingly common. However, they suffer from serious practical and ethical flaws which are difficult to tackle due to the fact that the inner workings of most models are closed to researchers. [Click here to see the full picture!]( [Data Science Facebook Groups You Should Join Today]( In brief: In the past, it was generally accepted that Facebook was for personal communications and LinkedIn for business. However, how people are using social media is changing and many are growing increasingly dependent on Facebook groups for information and to provide an important new forum for discussion. Being a data scientist can at times be rather solitary but groups can provide valuable networking opportunities alongside healthy debate and the sharing of industry news. This article by Analytics Insight introduces ten of the most useful Facebook groups in big data, data science, and ML. The groups vary from one which offers sharing material related to data management disciplines such as data, big data, analytics, data warehousing, data mining, business intelligence, and predictive analytics, to a platform used to share knowledge, problems, and solutions based on Big Data Microsoft HDInsight, Hadoop on Windows. Why this is important: This article is rather short and sweet but the groups listed are a great resource for people who work in or are merely interested in the field of data science. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, Dr. Thomas Wiecki, Core Developer of the PyMC Library and CEO of PyMC Labs, joins us for a masterclass in Bayesian statistics, which explains why it can be more powerful and interpretable than any other data modeling approach. --------------------------------------------------------------- What is the Data Science Insider? This email is a briefing of the week's most disruptive, interesting, and useful resources curated by the SuperDataScience team for Data Scientists who want to take their careers to the next level. Want to take your data science skills to the next level? Check out the [SuperDataScience platform]( and sign up for membership today! Know someone who would benefit from getting The Data Science Insider? Send them [this link to sign up.]( # # If you wish to stop receiving our emails or change your subscription options, please [Manage Your Subscription](
SuperDataScience Pty Ltd (ABN 91 617 928 131), 15 Macleay Crescent, Pacific Paradise, QLD 4564, Australia